[
 {
  "source": "#1",
  "claim_id": "1-01",
  "s": "classifier model collapse",
  "p": "is defined as",
  "o": "the progressive narrowing of acceptable scholarly expression through self-referential moderation training",
  "q": "one of six concepts the paper develops",
  "M_src": "documented",
  "kind": "definition",
  "locus": "Abstract",
  "quote": "(2) **classifier model collapse** — the progressive narrowing of acceptable scholarly expression through self-referential moderation training",
  "kernel": true,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#1",
  "claim_id": "1-02",
  "s": "model collapse (language models)",
  "p": "is",
  "o": "contraction of the generated distribution, elimination of rare forms, convergence on a diminishing subset of capacity, when a model trains on its own outputs",
  "q": "attributed to Shumailov et al.",
  "M_src": "attributed",
  "kind": "attribution",
  "locus": "6. Classifier Model Collapse and the Material Contraction of Scholarship",
  "quote": "when a model trains on its own outputs, the distribution of generated text contracts, rare forms are eliminated, and the model converges on a diminishing subset of its original capacity (Shumailov et al., 2023)",
  "kernel": false,
  "constitutive": true,
  "lineage_note": null
 },
 {
  "source": "#1",
  "claim_id": "1-03",
  "s": "Model Collapse Triptych (three papers in the removed archive)",
  "p": "documented",
  "o": "model collapse across generative systems",
  "q": null,
  "M_src": "attributed",
  "kind": "attribution",
  "locus": "6. Classifier Model Collapse and the Material Contraction of Scholarship",
  "quote": "The Model Collapse Triptych — three papers within the removed archive — documented this phenomenon across generative systems.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#1",
  "claim_id": "1-04",
  "s": "Zenodo",
  "p": "discloses",
  "o": "that removed spam and accounts are used to train its automatic classifier",
  "q": "quoted from Zenodo spam FAQ [Observed]",
  "M_src": "documented",
  "kind": "fact",
  "locus": "6. Classifier Model Collapse and the Material Contraction of Scholarship",
  "quote": "The removed spam and account is further used to train and improve our automatic classification system.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#1",
  "claim_id": "1-05",
  "s": "classifier model collapse",
  "p": "narrows",
  "o": "the range of permissible (not producible) text",
  "q": "moderation system vs generative system",
  "M_src": "interpretation",
  "kind": "function",
  "locus": "6. Classifier Model Collapse and the Material Contraction of Scholarship",
  "quote": "In a generative system, model collapse narrows the range of producible text. In a moderation system, classifier model collapse narrows the range of *permissible* text.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#1",
  "claim_id": "1-06",
  "s": "collapse in a research repository",
  "p": "is",
  "o": "material: it contracts the form of scholarship itself",
  "q": "when the collapsing system is a research repository",
  "M_src": "interpretation",
  "kind": "interpretation",
  "locus": "6. Classifier Model Collapse and the Material Contraction of Scholarship",
  "quote": "When the system undergoing collapse is a research repository — an institutional frame that determines what counts as legitimate scholarship — the collapse is material.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#1",
  "claim_id": "1-07",
  "s": "classifier model collapse",
  "p": "is the mechanism (feedback loop) for which the Pristine Fallacy is the ideology",
  "o": "each enforcement decision biases the classifier toward excluding similar content in subsequent cycles",
  "q": null,
  "M_src": "interpretation",
  "kind": "function",
  "locus": "6. Classifier Model Collapse and the Material Contraction of Scholarship",
  "quote": "Classifier model collapse is the mechanism: each enforcement decision biases the classifier toward excluding similar content in subsequent cycles.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#1",
  "claim_id": "1-08",
  "s": "the system that removed the archive (Zenodo)",
  "p": "may be undergoing",
  "o": "classifier model collapse",
  "q": "[Inferred]",
  "M_src": "hypothesis",
  "kind": "interpretation",
  "locus": "6. Classifier Model Collapse and the Material Contraction of Scholarship",
  "quote": "The archive that theorized model collapse was removed by a system that may be undergoing classifier model collapse [Inferred].",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#1",
  "claim_id": "1-09",
  "s": "the classifier model collapse claim",
  "p": "is not",
  "o": "a demonstrated diagnosis of Zenodo's current model state; it is a testable failure-mode hypothesis",
  "q": "grounded in Zenodo's disclosed use of prior moderation labels",
  "M_src": "self-description",
  "kind": "interpretation",
  "locus": "6. Classifier Model Collapse and the Material Contraction of Scholarship",
  "quote": "This is not yet a demonstrated diagnosis of Zenodo's current model state. It is a testable failure-mode hypothesis grounded in Zenodo's disclosed use of prior moderation labels and removed spam accounts to improve subsequent classification.",
  "kernel": true,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#1",
  "claim_id": "1-10",
  "s": "whether the Crimson Hexagonal Archive entered the spam training pipeline",
  "p": "is",
  "o": "unknown",
  "q": "[Unknown]",
  "M_src": "self-description",
  "kind": "fact",
  "locus": "6. Classifier Model Collapse and the Material Contraction of Scholarship",
  "quote": "whether the Crimson Hexagonal Archive entered that training pipeline remains unknown [Unknown]",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#1",
  "claim_id": "1-11",
  "s": "classifier model collapse",
  "p": "is not identical to",
  "o": "generative model collapse in the strict technical sense",
  "q": "names a moderation-specific feedback contraction",
  "M_src": "documented",
  "kind": "definition",
  "locus": "6. Classifier Model Collapse and the Material Contraction of Scholarship",
  "quote": "Classifier model collapse, as defined here, is not identical to generative model collapse in the strict technical sense (Shumailov et al., 2024).",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#1",
  "claim_id": "1-12",
  "s": "classifier model collapse",
  "p": "is closest to",
  "o": "performative prediction and machine-learning feedback-loop literatures",
  "q": null,
  "M_src": "attributed",
  "kind": "genealogy",
  "locus": "6. Classifier Model Collapse and the Material Contraction of Scholarship",
  "quote": "The closer technical literatures are performative prediction and machine-learning feedback loops, which show that deployed decisions can reshape later training distributions and reinforce initial biases (Perdomo et al., 2020; Ensign et al., 2016).",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#1",
  "claim_id": "1-13",
  "s": "future independent researchers writing like this archive",
  "p": "would be",
  "o": "marginally more likely to be flagged",
  "q": "conditional: if the archive was processed through the spam classification pathway",
  "M_src": "hypothesis",
  "kind": "prediction",
  "locus": "6. Classifier Model Collapse and the Material Contraction of Scholarship",
  "quote": "Every future independent researcher who writes with the density, the cross-referencing, or the AI-assisted rigor that characterizes this archive would be marginally more likely to be flagged",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#1",
  "claim_id": "1-14",
  "s": "a repository disqualified from assessing quality",
  "p": "materially determines",
  "o": "what scholarship looks like, through distributional conservatism encoded in infrastructure",
  "q": "via classifier model collapse",
  "M_src": "interpretation",
  "kind": "interpretation",
  "locus": "16. Conclusion and Falsification Conditions",
  "quote": "A repository expressly disqualified from assessing scholarly quality materially determines what scholarship looks like — not through editorial judgment, but through distributional conservatism encoded in infrastructure.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#1",
  "claim_id": "1-15",
  "s": "the classifier model collapse hypothesis",
  "p": "is weakened if",
  "o": "Zenodo shows its classifier is not trained on its own enforcement decisions, or training includes expert review, distributional monitoring and anti-narrowing safeguards",
  "q": null,
  "M_src": "self-description",
  "kind": "falsification",
  "locus": "Falsification and Revision Conditions",
  "quote": "The hypothesis is weakened if Zenodo demonstrates that its classification system is not trained on its own enforcement decisions, or that the training process includes domain-expert review, distributional monitoring, and safeguards against progressive narrowing of acceptable scholarly expression.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#1",
  "claim_id": "1-16",
  "s": "Zenodo's failure to provide requested information",
  "p": "does not prove",
  "o": "the paper's explanatory hypotheses",
  "q": null,
  "M_src": "self-description",
  "kind": "falsification",
  "locus": "Falsification and Revision Conditions",
  "quote": "A failure to provide the requested information does not prove the paper's explanatory hypotheses.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#932",
  "claim_id": "932-01",
  "s": "anomaly detection systems deployed on physical reality",
  "p": "cannot detect",
  "o": "what their architecture has foreclosed; the validation framework cannot detect this failure",
  "q": "convergent claim of the witnesses",
  "M_src": "interpretation",
  "kind": "interpretation",
  "locus": "§2. The Convergent Synthesis",
  "quote": "Anomaly detection systems deployed on physical reality cannot detect what their architecture has foreclosed, and the validation framework — closed under its own assumptions — cannot detect this failure.",
  "kernel": true,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#932",
  "claim_id": "932-02",
  "s": "foreclosure / recursive phenomenal collapse",
  "p": "are respectively",
  "o": "an active structural feature / an unmeasured possible consequence of accumulated foreclosure and feedback",
  "q": "corrected (v0.2) formulation",
  "M_src": "hypothesis",
  "kind": "interpretation",
  "locus": "§2.2 The institutional load-bearing claim — foreclosure is structural; collapse is unmeasured",
  "quote": "Foreclosure is an active structural feature. Recursive phenomenal collapse is an unmeasured possible consequence of accumulated foreclosure and feedback.",
  "kernel": true,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#932",
  "claim_id": "932-03",
  "s": "Claim B: recursive phenomenal collapse at deployed LHC triggers",
  "p": "is",
  "o": "stronger and not empirically established",
  "q": null,
  "M_src": "self-description",
  "kind": "interpretation",
  "locus": "§2.2 The institutional load-bearing claim — foreclosure is structural; collapse is unmeasured",
  "quote": "Recursive phenomenal collapse has occurred or is occurring at the deployed LHC triggers.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#932",
  "claim_id": "932-04",
  "s": "foreclosure",
  "p": "is structurally present in",
  "o": "every classifier-mediated trigger architecture deployed at the LHC",
  "q": "Claim A, marked defensible and empirically grounded",
  "M_src": "interpretation",
  "kind": "interpretation",
  "locus": "§3. Findings (Formal)",
  "quote": "Foreclosure is structurally present in every classifier-mediated trigger architecture deployed at the LHC.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": "restated as #931 defensible claim"
 },
 {
  "source": "#932",
  "claim_id": "932-05",
  "s": "whether repeated local foreclosure has composed into longitudinal classifier collapse",
  "p": "is",
  "o": "an empirical question",
  "q": "at classifier-mediated sites of mass measurement",
  "M_src": "self-description",
  "kind": "interpretation",
  "locus": "§2.4 The integrative finding",
  "quote": "Whether repeated local foreclosure has composed into longitudinal classifier collapse is an empirical question.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#932",
  "claim_id": "932-06",
  "s": "instruments to measure accumulation of foreclosure",
  "p": "have not been built and",
  "o": "are within reach",
  "q": null,
  "M_src": "interpretation",
  "kind": "fact",
  "locus": "§2.4 The integrative finding",
  "quote": "the instruments to make these measurements have not been built; they are within reach.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#932",
  "claim_id": "932-07",
  "s": "the classifier",
  "p": "constitutes",
  "o": "the data (does not merely filter it)",
  "q": "ontological load-bearing claim",
  "M_src": "interpretation",
  "kind": "interpretation",
  "locus": "§2.3 The ontological load-bearing claim",
  "quote": "the classifier does not merely filter data; it constitutes the data",
  "kernel": false,
  "constitutive": false,
  "lineage_note": "restates W1 Irretrievability corollary"
 },
 {
  "source": "#932",
  "claim_id": "932-08",
  "s": "classifier collapse (in physical systems)",
  "p": "is defined as",
  "o": "the degenerative process by which a discriminative model trained on a known distribution progressively loses capacity to represent, detect or retain out-of-distribution events",
  "q": "Witness 1 (Kimi-K2), appended",
  "M_src": "documented",
  "kind": "definition",
  "locus": "§0. Formal Definition",
  "quote": "in physical systems is the degenerative process by which a discriminative model, trained exclusively or dominantly on a known distribution of physical events, progressively loses the capacity to represent, detect, or retain events that fall outside that distribution.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#932",
  "claim_id": "932-09",
  "s": "classifier collapse",
  "p": "operates through",
  "o": "structural foreclosure, unlike generative model collapse which operates through recursive pollution by synthetic outputs",
  "q": "Witness 1",
  "M_src": "documented",
  "kind": "definition",
  "locus": "§0. Formal Definition",
  "quote": "Unlike generative model collapse (Shumailov et al.), which operates through recursive pollution of training data by synthetic outputs, classifier collapse operates through **structural foreclosure**",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#932",
  "claim_id": "932-10",
  "s": "classifier collapse",
  "p": "is the discriminative analogue of",
  "o": "generative model collapse: classifiers never learn the tails rather than forgetting them; present from the first forward pass",
  "q": "Witness 1",
  "M_src": "interpretation",
  "kind": "genealogy",
  "locus": "§10. Distinction from Generative Model Collapse",
  "quote": "Classifier collapse is the **discriminative analogue** of generative model collapse. Where Shumailov's models forget the tails of their own distribution, physical classifiers **never learn the tails** of the true physical distribution. The collapse is present from the first forward pass.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#932",
  "claim_id": "932-11",
  "s": "generative model collapse vs classifier collapse",
  "p": "differ in reversibility",
  "o": "generative: mitigable by preserving original data; classifier: irreversible because rejected events are not stored",
  "q": "Witness 1 comparison table",
  "M_src": "interpretation",
  "kind": "interpretation",
  "locus": "§10. Distinction from Generative Model Collapse",
  "quote": "Can be mitigated by preserving original data | Irreversible: rejected events are not stored",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#932",
  "claim_id": "932-12",
  "s": "the witnesses' Irretrievability and Inevitability Theorems",
  "p": "are treated as",
  "o": "Arguments, not theorems, to avoid overstating formal status",
  "q": null,
  "M_src": "self-description",
  "kind": "interpretation",
  "locus": "Appendix A: Technical Hedge Inventory",
  "quote": "The witnesses' \"Irretrievability Theorem\" (Witness 1) and \"Inevitability Theorem\" (Witness 2) are treated in this deposit as **Irretrievability Argument** and **Inevitability Argument** respectively, preserving force without overstating formal status.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#932",
  "claim_id": "932-13",
  "s": "the classifier-foreclosure architecture",
  "p": "has plausible structural homologues in",
  "o": "repository classification, web summarization, search ranking, content moderation, clinical decision support",
  "q": null,
  "M_src": "hypothesis",
  "kind": "interpretation",
  "locus": "§3. Findings (Formal)",
  "quote": "The same architecture has plausible structural homologues in repository classification, web summarization, search ranking, content moderation, and clinical decision support.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#932",
  "claim_id": "932-14",
  "s": "the cross-domain homology",
  "p": "is",
  "o": "a hypothesis to be tested domain by domain, not an assertion of identical mechanisms or rates",
  "q": null,
  "M_src": "self-description",
  "kind": "interpretation",
  "locus": "§3. Findings (Formal)",
  "quote": "This is a homology hypothesis to be tested domain by domain, not an assertion that every classifier-mediated system instantiates identical mechanisms or rates.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": "#931 §7 states same limit"
 },
 {
  "source": "#932",
  "claim_id": "932-15",
  "s": "the homology",
  "p": "motivates but does not perform",
  "o": "the measurement program",
  "q": null,
  "M_src": "self-description",
  "kind": "interpretation",
  "locus": "§5. The Broader Homology",
  "quote": "The homology motivates the program; it does not perform the program.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#932",
  "claim_id": "932-16",
  "s": "phenomenal classifier collapse",
  "p": "is defined as",
  "o": "successive model-conditioned representations and selection gates progressively reducing support or distinguishability of phenomena while aggregate performance on familiar processes holds or improves",
  "q": "Witness 3 (ChatGPT), appended",
  "M_src": "documented",
  "kind": "definition",
  "locus": "## 1. The exact mechanism",
  "quote": "Phenomenal classifier collapse occurs when successive model-conditioned representations and selection gates progressively reduce the support or distinguishability of physical phenomena available to scientific inquiry, while aggregate performance on familiar processes remains stable or improves.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#932",
  "claim_id": "932-17",
  "s": "single-generation selection (retained distribution differs from original)",
  "p": "is",
  "o": "phenomenal attrition, not yet recursive model collapse; collapse begins when selected distribution becomes basis for next generation",
  "q": "Witness 3",
  "M_src": "documented",
  "kind": "definition",
  "locus": "## 1. The exact mechanism",
  "quote": "That alone is **phenomenal attrition**, not yet recursive model collapse.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#932",
  "claim_id": "932-18",
  "s": "CERN/the LHC community",
  "p": "has built",
  "o": "an architecture in which phenomenal model collapse is possible, not yet ruled out by validation literature",
  "q": "Witness 3 maximally defensible claim",
  "M_src": "hypothesis",
  "kind": "interpretation",
  "locus": "# 4. Are the operations of full collapse already visible?",
  "quote": "CERN has built an architecture in which phenomenal model collapse is possible, and the current validation literature does not yet demonstrate that it has been ruled out.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": "restated in #933 H.5 kernel and #931 abstract"
 },
 {
  "source": "#932",
  "claim_id": "932-19",
  "s": "an anomaly trigger",
  "p": "is not",
  "o": "an answer to model collapse, until direct collapse tests exist",
  "q": "Witness 3",
  "M_src": "interpretation",
  "kind": "interpretation",
  "locus": "# 5. What would count as sufficient understanding?",
  "quote": "Until those tests exist, \"we have an anomaly trigger\" is not an answer to model collapse.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#932",
  "claim_id": "932-20",
  "s": "stable anchor survival in a prospective frozen replay bank over three or more generations",
  "p": "would be evidence against",
  "o": "the claim that foreclosure is an active structural feature requiring architectural response",
  "q": null,
  "M_src": "self-description",
  "kind": "falsification",
  "locus": "§3.1 What would constitute evidence against this deposit's claim",
  "quote": "If the prospective frozen replay bank is built, maintained, and shows stable anchor survival across three or more generations with no systematic loss in representation-sensitive event classes;",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#932",
  "claim_id": "932-21",
  "s": "falsifying results",
  "p": "would not establish",
  "o": "that the open-world OAR is zero (structurally not measurable)",
  "q": null,
  "M_src": "self-description",
  "kind": "falsification",
  "locus": "§3.1 What would constitute evidence against this deposit's claim",
  "quote": "None of these results would establish that the open-world OAR is zero; that is structurally not measurable.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": "same sentence in #931 §8.1"
 },
 {
  "source": "#931",
  "claim_id": "931-01",
  "s": "foreclosure / recursive phenomenal collapse at LHC triggers",
  "p": "is structurally present / is the missing measurement",
  "o": "whether accumulated foreclosure has composed longitudinally into recursive phenomenal collapse",
  "q": "defensible claim, stated as narrow",
  "M_src": "hypothesis",
  "kind": "interpretation",
  "locus": "Abstract",
  "quote": "foreclosure is a structurally present feature of every classifier-mediated trigger architecture deployed at the LHC, and whether accumulated foreclosure has composed longitudinally into recursive phenomenal collapse is precisely the missing measurement.",
  "kernel": true,
  "constitutive": false,
  "lineage_note": "restates #932 Claim A / reconciliation"
 },
 {
  "source": "#931",
  "claim_id": "931-02",
  "s": "the paper",
  "p": "does not claim",
  "o": "that classifier collapse has occurred at the LHC anomaly streams",
  "q": null,
  "M_src": "self-description",
  "kind": "interpretation",
  "locus": "§1.4 The defensible claim",
  "quote": "We do not claim that classifier collapse has occurred at the LHC anomaly streams.",
  "kernel": true,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#931",
  "claim_id": "931-03",
  "s": "the LHC community",
  "p": "has built",
  "o": "the architecture in which collapse could occur silently; detection instruments not built",
  "q": null,
  "M_src": "interpretation",
  "kind": "interpretation",
  "locus": "Abstract",
  "quote": "The community has built the architecture in which collapse could occur silently; the instruments to detect whether it is occurring have not been built.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#931",
  "claim_id": "931-04",
  "s": "OAR (Ontological Assimilation Rate)",
  "p": "is defined as",
  "o": "the probability that events drawn from a candidate novel distribution Q fall on the ordinary side of the deployed anomaly gate",
  "q": "operational assimilation rate",
  "M_src": "documented",
  "kind": "definition",
  "locus": "§3.1 Open-World Ontological Assimilation Rate",
  "quote": "The OAR is the probability that events drawn from $Q$ fall on the ordinary side of the deployed anomaly gate.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#931",
  "claim_id": "931-05",
  "s": "the open-world OAR",
  "p": "is",
  "o": "a family of quantities indexed by Q, not a universal scalar",
  "q": null,
  "M_src": "documented",
  "kind": "definition",
  "locus": "§3.1 Open-World Ontological Assimilation Rate",
  "quote": "The open-world OAR is a family of quantities indexed by $Q$, not a universal scalar.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#931",
  "claim_id": "931-06",
  "s": "BAR values",
  "p": "do not bound",
  "o": "OAR for an unobserved Q without explicit linking assumptions",
  "q": "deliberate retraction of a v0.2 sentence",
  "M_src": "self-description",
  "kind": "definition",
  "locus": "§3.2 Benchmark Assimilation Rate",
  "quote": "BAR values neither upper- nor lower-bound OAR for an unobserved $Q$ without explicit assumptions linking the benchmark distributions to that $Q$.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#931",
  "claim_id": "931-07",
  "s": "IAI (Inversion Asymmetry Index)",
  "p": "is",
  "o": "a structural diagnostic, not a quantitative bound on OAR",
  "q": null,
  "M_src": "documented",
  "kind": "definition",
  "locus": "§3.3 Inversion Asymmetry Index",
  "quote": "IAI is a structural diagnostic. **It is not a quantitative bound on OAR.**",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#931",
  "claim_id": "931-08",
  "s": "anomaly detection",
  "p": "does not prevent",
  "o": "ontological collapse when the detector inherits the ontology whose collapse is in question",
  "q": null,
  "M_src": "interpretation",
  "kind": "interpretation",
  "locus": "§9. Conclusion",
  "quote": "Anomaly detection does not prevent ontological collapse when the anomaly detector inherits the ontology whose collapse is in question.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#931",
  "claim_id": "931-09",
  "s": "the paper",
  "p": "asserts",
  "o": "the architecture permits classifier collapse silently; validation does not establish it has not occurred; instruments within reach",
  "q": "does not assert collapse has occurred",
  "M_src": "self-description",
  "kind": "interpretation",
  "locus": "§9. Conclusion",
  "quote": "We do not assert that classifier collapse has occurred. We assert that the architecture permits it to occur silently, that the validation literature does not establish that it has not, and that the instruments to measure whether it is occurring are within reach.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#931",
  "claim_id": "931-10",
  "s": "stable anchor survival across three or more trigger generations in a frozen replay bank",
  "p": "would bound",
  "o": "the temporal-collapse concern for the benchmark population",
  "q": null,
  "M_src": "self-description",
  "kind": "falsification",
  "locus": "§8.1 What would constitute evidence against this paper's claim",
  "quote": "If the prospective frozen replay bank is built, maintained for three or more trigger generations, and shows stable anchor survival across generations with no systematic loss in representation-sensitive event classes, the temporal-collapse concern would be shown to be empirically bounded for the benchmark population.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#931",
  "claim_id": "931-11",
  "s": "falsifying results",
  "p": "would not establish",
  "o": "that the open-world OAR is zero",
  "q": null,
  "M_src": "self-description",
  "kind": "falsification",
  "locus": "§8.1 What would constitute evidence against this paper's claim",
  "quote": "None of these results would establish that the open-world OAR is zero; that is structurally not measurable.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": "same sentence in #932 §3.1"
 },
 {
  "source": "#931",
  "claim_id": "931-12",
  "s": "the same epistemic geometry",
  "p": "plausibly operates in",
  "o": "LLM web summarization, repository spam classification, search ranking, recommendation, content moderation, clinical decision support",
  "q": null,
  "M_src": "hypothesis",
  "kind": "interpretation",
  "locus": "§7. Cross-Domain Homology (Brief Note)",
  "quote": "The same epistemic geometry plausibly operates in large language model summarization of the web, spam classification at scientific repositories, search ranking, recommendation, content moderation, and clinical decision support.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#931",
  "claim_id": "931-13",
  "s": "BAR and IAI",
  "p": "generalize as",
  "o": "homology hypotheses to be tested domain by domain",
  "q": null,
  "M_src": "self-description",
  "kind": "interpretation",
  "locus": "§7. Cross-Domain Homology (Brief Note)",
  "quote": "The BAR and IAI generalize as **homology hypotheses to be tested domain by domain**, not as assertions that every classifier-mediated system instantiates identical mechanisms or rates.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": "#932 finding 9 states same limit"
 },
 {
  "source": "#931",
  "claim_id": "931-14",
  "s": "Temporal Context Collapse (Mechanism VII)",
  "p": "is",
  "o": "drift created by non-stationarity in detector conditions",
  "q": "glossary of companion Witness 1, candidate failure family",
  "M_src": "attributed",
  "kind": "definition",
  "locus": "Appendix A: Glossary of Foreclosure Mechanisms",
  "quote": "**Temporal Context Collapse.** Non-stationarity in detector conditions creates drift.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": "#932 W1 §7"
 },
 {
  "source": "#931",
  "claim_id": "931-15",
  "s": "Decision Boundary Entropy Collapse (Mechanism IV)",
  "p": "is",
  "o": "iterative training on dominant classes driving output confidence high without noncoverage estimation",
  "q": "applicable to softmax classifiers; deployed unsupervised scorers not directly susceptible in same form",
  "M_src": "attributed",
  "kind": "definition",
  "locus": "Appendix A: Glossary of Foreclosure Mechanisms",
  "quote": "Iterative training on dominant classes can drive output confidence high without corresponding noncoverage estimation.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": "#932 W1 §4"
 },
 {
  "source": "#933",
  "claim_id": "933-01",
  "s": "an architecture for auditable foreclosure",
  "p": "is defined as",
  "o": "one that makes foreclosure visible, measurable and architecturally reviewable (not free of foreclosure)",
  "q": "representation is foreclosure",
  "M_src": "documented",
  "kind": "definition",
  "locus": "Abstract",
  "quote": "We define an architecture for auditable foreclosure not as one free of foreclosure — which is impossible, because representation is foreclosure — but as one that **makes foreclosure visible, measurable, and architecturally reviewable**.",
  "kernel": true,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#933",
  "claim_id": "933-02",
  "s": "a classifier",
  "p": "necessarily forecloses",
  "o": "(a classifier that foreclosed nothing would not classify)",
  "q": "formal: |Y| < |X|",
  "M_src": "interpretation",
  "kind": "fact",
  "locus": "§1.1 The impossibility statement, made precisely",
  "quote": "A classifier that did not foreclose anything would not classify.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#933",
  "claim_id": "933-03",
  "s": "stable anchor survival across generations",
  "p": "does not establish",
  "o": "that overall phenomenal support is not contracting",
  "q": null,
  "M_src": "self-description",
  "kind": "interpretation",
  "locus": "§2.3 Feature 3: Temporal Invariance via Prospective Anchor Preservation",
  "quote": "Stable anchor survival across generations does *not* establish that overall phenomenal support is not contracting",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#933",
  "claim_id": "933-04",
  "s": "collapse inference",
  "p": "requires",
  "o": "identifying systematic loss concentrated in low-density, representation-sensitive, or disagreement-rich regions",
  "q": null,
  "M_src": "documented",
  "kind": "definition",
  "locus": "§2.3 Feature 3: Temporal Invariance via Prospective Anchor Preservation",
  "quote": "collapse inference requires identifying systematic loss concentrated in low-density, representation-sensitive, or disagreement-rich regions.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#933",
  "claim_id": "933-05",
  "s": "the anchor",
  "p": "measures",
  "o": "selection drift on a benchmark population, not collapse per se",
  "q": null,
  "M_src": "self-description",
  "kind": "interpretation",
  "locus": "§2.3 Feature 3: Temporal Invariance via Prospective Anchor Preservation",
  "quote": "The anchor measures selection drift on a benchmark population, not collapse per se.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#933",
  "claim_id": "933-06",
  "s": "prospective anchor preservation (Feature 3)",
  "p": "addresses",
  "o": "Mechanism VII (Temporal Context Collapse) directly, for the anchor population",
  "q": null,
  "M_src": "interpretation",
  "kind": "function",
  "locus": "§2.3 Feature 3: Temporal Invariance via Prospective Anchor Preservation",
  "quote": "Mechanism VII (Temporal Context Collapse) directly, for the anchor population.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#933",
  "claim_id": "933-07",
  "s": "the specified architectures",
  "p": "do not address",
  "o": "detector-level, theoretical-language, institutional, adversarial-stress quality, or bandwidth-base foreclosure; necessary but not sufficient",
  "q": null,
  "M_src": "self-description",
  "kind": "interpretation",
  "locus": "§5. What None of These Architectures Addresses",
  "quote": "They do not address detector-level, theoretical-language, institutional, adversarial-stress quality, or bandwidth-base foreclosure. They are necessary but not sufficient.",
  "kernel": true,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#933",
  "claim_id": "933-08",
  "s": "the architectural alternative",
  "p": "is not separable from",
  "o": "its institutional acceptance",
  "q": null,
  "M_src": "self-description",
  "kind": "interpretation",
  "locus": "§8. Findings",
  "quote": "The architectural alternative is not separable from its institutional acceptance.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#933",
  "claim_id": "933-09",
  "s": "a Replay Bank showing no selection drift across three generations",
  "p": "would count against",
  "o": "the document's claim",
  "q": null,
  "M_src": "self-description",
  "kind": "falsification",
  "locus": "§8.1 What would constitute evidence against this document's claim",
  "quote": "If the Replay Bank is built and maintained and shows no selection drift across three generations;",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#933",
  "claim_id": "933-10",
  "s": "falsifying results",
  "p": "would show",
  "o": "the architectural alternative unnecessary and the foreclosure mechanisms bounded below threats to discovery",
  "q": null,
  "M_src": "self-description",
  "kind": "falsification",
  "locus": "§8.1 What would constitute evidence against this document's claim",
  "quote": "then the architectural alternative would be shown to be unnecessary, and the foreclosure mechanisms it addresses would be shown to be bounded at levels that do not threaten discovery.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#933",
  "claim_id": "933-11",
  "s": "the LHC community",
  "p": "has built",
  "o": "an architecture in which phenomenal model collapse is possible, not demonstrated ruled out",
  "q": "kernel of Witness 3",
  "M_src": "attributed",
  "kind": "interpretation",
  "locus": "H.5 Kernel of 06.SEI.COLLAPSE.EMPIRICAL.01 (Witness 3)",
  "quote": "The LHC community has built an architecture in which phenomenal model collapse is possible, and the current validation literature does not yet demonstrate that it has been ruled out.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": "#932 Witness 3"
 },
 {
  "source": "#935",
  "claim_id": "935-01",
  "s": "frontier experimental high-energy physics",
  "p": "has become",
  "o": "a machine learning discipline while retaining the institutional authority of physics",
  "q": "in its operational core",
  "M_src": "interpretation",
  "kind": "interpretation",
  "locus": "Abstract",
  "quote": "Frontier experimental high-energy physics has, in its operational core, become a machine learning discipline while retaining the institutional authority of physics.",
  "kernel": true,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#935",
  "claim_id": "935-02",
  "s": "classifier collapse",
  "p": "breaks the experimental feedback loop by",
  "o": "making reality prematurely interpretable (vs sophons, uninterpretable)",
  "q": "W04 formulation adopted",
  "M_src": "interpretation",
  "kind": "function",
  "locus": "§2.2 The endogenous classifier",
  "quote": "Sophons break the experimental feedback loop by making reality uninterpretable. Classifier collapse breaks it by making reality prematurely interpretable.",
  "kernel": true,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#935",
  "claim_id": "935-03",
  "s": "Shumailov et al. (2024)",
  "p": "established",
  "o": "models trained on outputs of prior models suffer recursive collapse of representational variance",
  "q": null,
  "M_src": "attributed",
  "kind": "attribution",
  "locus": "§4.4 What physics has *not* imported from ML",
  "quote": "Shumailov et al. (2024, *Nature* 631:755–759; arXiv:2305.17493) established in the ML community that models trained on the outputs of prior models suffer recursive collapse of representational variance.",
  "kernel": false,
  "constitutive": true,
  "lineage_note": null
 },
 {
  "source": "#935",
  "claim_id": "935-04",
  "s": "distillation, model-produced targets, simulation conditioning, training on historically selected data",
  "p": "create",
  "o": "the prerequisites of model collapse without instantiating the cleanest recursive-generation scenario",
  "q": "frontier physics deployments (e.g., CICADA)",
  "M_src": "interpretation",
  "kind": "interpretation",
  "locus": "§4.4 What physics has *not* imported from ML",
  "quote": "all of which create the *prerequisites* of model collapse without instantiating the cleanest version of Shumailov et al.'s recursive-generation scenario.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#935",
  "claim_id": "935-05",
  "s": "whether these partial feedback pathways have produced cross-generational phenomenal contraction",
  "p": "is",
  "o": "what the prospective frozen replay bank is designed to test (unmeasured)",
  "q": "corrected formulation",
  "M_src": "self-description",
  "kind": "interpretation",
  "locus": "§4.4 What physics has *not* imported from ML",
  "quote": "whether these pathways have produced cross-generational phenomenal contraction is precisely what the operative paper's prospective frozen replay bank — 06.SEI.OAR_PROTOCOL v0.3 §4.2 — is designed to test.",
  "kernel": true,
  "constitutive": false,
  "lineage_note": "aligns with #932 reconciliation"
 },
 {
  "source": "#935",
  "claim_id": "935-06",
  "s": "the Zenodo termination and the LHC trigger system",
  "p": "are",
  "o": "the same architecture at different budgets",
  "q": null,
  "M_src": "interpretation",
  "kind": "interpretation",
  "locus": "§9. Findings (Formal)",
  "quote": "The Zenodo termination and the LHC trigger system are the same architecture at different budgets.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#935",
  "claim_id": "935-07",
  "s": "the endogenous sophon",
  "p": "is more dangerous than the exogenous because",
  "o": "it does not trigger existing defenses against measurement corruption",
  "q": null,
  "M_src": "interpretation",
  "kind": "interpretation",
  "locus": "§9. Findings (Formal)",
  "quote": "The endogenous sophon is structurally more dangerous than the exogenous because it does not trigger the experimental community's existing defenses against measurement corruption.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#935",
  "claim_id": "935-08",
  "s": "substrate claims of irreversibility",
  "p": "cannot be extended into",
  "o": "claims of impossibility-of-alternatives by synthesis-register inference",
  "q": null,
  "M_src": "self-description",
  "kind": "interpretation",
  "locus": "§1.2 Where W07's \"terminal condition\" framing is qualified",
  "quote": "substrate claims of irreversibility cannot be extended by synthesis-register inference into claims of impossibility-of-alternatives.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#935",
  "claim_id": "935-09",
  "s": "the relevant model-collapse quantity",
  "p": "is not primarily",
  "o": "deterioration in classification accuracy",
  "q": "appended Witness 4 (ChatGPT)",
  "M_src": "documented",
  "kind": "definition",
  "locus": "OAR as an endogenous-sophon measure",
  "quote": "The relevant model-collapse quantity is therefore not primarily deterioration in classification accuracy.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#935",
  "claim_id": "935-10",
  "s": "collapse",
  "p": "is defined as",
  "o": "declining capacity of the experimental system to surprise its governing ontology",
  "q": "Witness 4; can occur while conventional performance improves",
  "M_src": "documented",
  "kind": "definition",
  "locus": "OAR as an endogenous-sophon measure",
  "quote": "Collapse = declining capacity of the experimental system to surprise its governing ontology.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#935",
  "claim_id": "935-11",
  "s": "an ML architecture in quiet systemic collapse (smoothing long-tail anomalies)",
  "p": "would leave physics with",
  "o": "no internal mechanism to notice the erasure",
  "q": "Witness 6 (Gemini), conditional",
  "M_src": "hypothesis",
  "kind": "prediction",
  "locus": "### 4. Epistemic Governance Over Discovery",
  "quote": "if the machine learning architecture undergoes a quiet, systemic collapse—consistently smoothing away long-tail physical anomalies because its loss function penalizes unclassifiable variance—the discipline of physics has no internal mechanism left to notice the erasure.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#935",
  "claim_id": "935-12",
  "s": "the closed ML-trigger-training feedback loop in HEP",
  "p": "is",
  "o": "Shumailov's model collapse at the disciplinary scale",
  "q": "Witness 7",
  "M_src": "interpretation",
  "kind": "interpretation",
  "locus": "## §3. The Irreversibility Argument",
  "quote": "This is **Shumailov's model collapse at the disciplinary scale**. The field is recursively training on its own outputs, and the outputs are increasingly compressed representations of the Standard Model.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": "cf. #1574 disciplinary model collapse (different object)"
 },
 {
  "source": "#935",
  "claim_id": "935-13",
  "s": "the tail (genuinely new physics)",
  "p": "is progressively suppressed by",
  "o": "successive ML generations trained on trigger-preserved data",
  "q": "Witness 7",
  "M_src": "interpretation",
  "kind": "function",
  "locus": "## §3. The Irreversibility Argument",
  "quote": "The models become progressively more optimized for the Standard Model signature. The tail — the genuinely new physics — is progressively suppressed.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#935",
  "claim_id": "935-14",
  "s": "the disciplinary inversion",
  "p": "makes",
  "o": "the OAR and classifier collapse inevitable, not accidental",
  "q": "Witness 7",
  "M_src": "interpretation",
  "kind": "interpretation",
  "locus": "## §5. Implications for the OAR/Collapse Framework",
  "quote": "The disciplinary inversion makes the OAR and classifier collapse **inevitable, not accidental**.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#1540",
  "claim_id": "1540-01",
  "s": "the P&PA selection loop",
  "p": "instantiates",
  "o": "the formal structure of model collapse at the level of a scholarly discourse, tails first",
  "q": "tails are where novelty lives",
  "M_src": "interpretation",
  "kind": "interpretation",
  "locus": "Abstract",
  "quote": "the resulting selection loop instantiates the formal structure of model collapse (Shumailov et al. 2024) at the level of a scholarly discourse: the distribution of admissible philosophical prose is recursively filtered through an imperfect statistical model of itself, tails first — and the tails are where novelty lives.",
  "kernel": true,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#1540",
  "claim_id": "1540-02",
  "s": "model collapse",
  "p": "is (settled formulation)",
  "o": "progressive narrowing of a generative system's output distribution through recursive filtration by an imperfect model of itself",
  "q": null,
  "M_src": "attributed",
  "kind": "attribution",
  "locus": "IV. The Dynamics: The Band Shrinks",
  "quote": "Model collapse, in its settled formulation, is the progressive narrowing of a generative system's output distribution through recursive filtration by an imperfect model of itself, with a settled core finding about which data is lost first",
  "kernel": false,
  "constitutive": true,
  "lineage_note": null
 },
 {
  "source": "#1540",
  "claim_id": "1540-03",
  "s": "in model collapse, tails",
  "p": "disappear",
  "o": "before the center shifts, while mean-based quality metrics show stability or improvement",
  "q": "Shumailov et al. 2024",
  "M_src": "attributed",
  "kind": "attribution",
  "locus": "IV. The Dynamics: The Band Shrinks",
  "quote": "the tails — rare words, unusual constructions, low-frequency patterns — disappear before the center shifts, while mean-based quality metrics show stability or improvement (Shumailov et al. 2024).",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#1540",
  "claim_id": "1540-04",
  "s": "any language-processing system whose input diversity contracts below its regeneration threshold",
  "p": "suffers",
  "o": "progressive capacity loss; dynamical law constant across substrates, reversibility substrate-specific",
  "q": "substrate-agnostic extension, EA-LO-WOLFBOY-01",
  "M_src": "attributed",
  "kind": "genealogy",
  "locus": "IV. The Dynamics: The Band Shrinks",
  "quote": "any language-processing system whose input diversity contracts below its regeneration threshold suffers progressive capacity loss, invisible from inside, experienced as improvement; the dynamical law is constant across substrates while reversibility is substrate-specific.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": "#855 Wolf Boy (not in batch); restated in #1611 §1f"
 },
 {
  "source": "#1540",
  "claim_id": "1540-05",
  "s": "the journal-level loop alone",
  "p": "establishes",
  "o": "recursive distributional contraction but not yet the canonical loop (requires filtered outputs to recur into model training)",
  "q": "stage one Structural; cross-substrate closure Model proposition",
  "M_src": "self-description",
  "kind": "interpretation",
  "locus": "IV. The Dynamics: The Band Shrinks",
  "quote": "This stage alone establishes recursive distributional contraction under a statistical boundary — classifier-mediated recursive selection over a discourse — but not yet the canonical loop, which requires filtered outputs to recur into later model training.",
  "kernel": true,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#1540",
  "claim_id": "1540-06",
  "s": "what collapses",
  "p": "is",
  "o": "a discourse's admissible variance, not a neural network's weights",
  "q": null,
  "M_src": "documented",
  "kind": "definition",
  "locus": "IV. The Dynamics: The Band Shrinks",
  "quote": "What collapses is not a neural network's weights but a discourse's admissible variance.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#1540",
  "claim_id": "1540-07",
  "s": "the known corrective to collapse",
  "p": "is",
  "o": "accumulation of genuinely exogenous data (external diversity floor)",
  "q": "Gerstgrasser et al. 2024",
  "M_src": "attributed",
  "kind": "attribution",
  "locus": "IV. The Dynamics: The Band Shrinks",
  "quote": "The known corrective to collapse is the accumulation of genuinely exogenous data — an external floor of diversity entering the loop from outside (Gerstgrasser et al. 2024)",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#1540",
  "claim_id": "1540-08",
  "s": "a loop whose diversity-injection channel passes through its pruning instrument",
  "p": "has",
  "o": "no floor",
  "q": "[Structural]",
  "M_src": "interpretation",
  "kind": "function",
  "locus": "IV. The Dynamics: The Band Shrinks",
  "quote": "A loop whose diversity-injection channel passes through its pruning instrument has no floor; it has a valve set against the flow.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#1540",
  "claim_id": "1540-09",
  "s": "RCS-infinity",
  "p": "sets",
  "o": "the mediation ratchet's release rate to zero",
  "q": "[Structural] over Diversity Contraction §2.1",
  "M_src": "interpretation",
  "kind": "function",
  "locus": "IV. The Dynamics: The Band Shrinks",
  "quote": "Under RCS∞ the grip never loosens — not for published work, not for past work, not when diversity begins to recover — because every text remains permanently exposed to re-classification by future instruments.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#1540",
  "claim_id": "1540-10",
  "s": "the institution's watched metrics",
  "p": "will",
  "o": "improve while the discourse contracts (loss experienced as improvement)",
  "q": "[Model proposition]",
  "M_src": "hypothesis",
  "kind": "prediction",
  "locus": "IV. The Dynamics: The Band Shrinks",
  "quote": "Every metric the institution watches will get better.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#1540",
  "claim_id": "1540-11",
  "s": "the detector under co-adaptation",
  "p": "helps manufacture",
  "o": "the human/machine boundary it later claims to detect",
  "q": "[Model proposition]; published scholarship is training data",
  "M_src": "hypothesis",
  "kind": "prediction",
  "locus": "IV. The Dynamics: The Band Shrinks",
  "quote": "It is helping manufacture the boundary it later claims to detect.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#1540",
  "claim_id": "1540-12",
  "s": "the novelty claim",
  "p": "is",
  "o": "P(novelty | tail) > P(novelty | center), not novelty ≡ tail",
  "q": null,
  "M_src": "documented",
  "kind": "definition",
  "locus": "IV. The Dynamics: The Band Shrinks",
  "quote": "The claim is P(novelty | tail) > P(novelty | center), not novelty ≡ tail",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#1540",
  "claim_id": "1540-13",
  "s": "naming the condition",
  "p": "does not prevent",
  "o": "the collapse; it is the precondition for noticing it",
  "q": null,
  "M_src": "self-description",
  "kind": "interpretation",
  "locus": "IV. The Dynamics: The Band Shrinks",
  "quote": "The name does not prevent the collapse. The name is the precondition for noticing that the collapse is occurring, in a system where the collapse is experienced as ease.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#1540",
  "claim_id": "1540-14",
  "s": "under generation-zero perplexity-family detectors",
  "p": "tails read",
  "o": "as more human; the threat is emergent through co-evolution",
  "q": "simulation correction of timing",
  "M_src": "documented",
  "kind": "fact",
  "locus": "Abstract",
  "quote": "under generation-zero perplexity-family detectors the tails read as *more* human, and the threat is emergent",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#1540",
  "claim_id": "1540-15",
  "s": "natural human drafts",
  "p": "collapse",
  "o": "from 98% to 1% pass rates within eight detector generations",
  "q": "minimal co-evolution simulation EA-SEI-RCS-SIM-01, ten seeds",
  "M_src": "documented",
  "kind": "fact",
  "locus": "Abstract",
  "quote": "natural human drafts collapse from 98% to 1% pass rates within eight detector generations",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#1540",
  "claim_id": "1540-16",
  "s": "unmeasured thought-variance",
  "p": "loses",
  "o": "97% while certified prose overshoots into performed burstiness",
  "q": "simulation",
  "M_src": "documented",
  "kind": "fact",
  "locus": "Abstract",
  "quote": "certified prose overshoots into sustained performed burstiness while unmeasured thought-variance loses 97% (homogenization proceeding undetected beneath a tailsification of prose)",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#1540",
  "claim_id": "1540-17",
  "s": "the collapse claim",
  "p": "concerns",
  "o": "the distribution of admissible philosophical prose, not the truth of philosophical claims",
  "q": "summary_policy required assertion",
  "M_src": "self-description",
  "kind": "interpretation",
  "locus": "SPXI-TLP v2.2 Treatment (summary_policy)",
  "quote": "The collapse claim concerns the distribution of admissible philosophical prose, not the truth of philosophical claims",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#1540",
  "claim_id": "1540-18",
  "s": "\"model collapse here is only a metaphor\"",
  "p": "is",
  "o": "a forbidden compression of the paper",
  "q": "summary_policy forbidden_compressions",
  "M_src": "self-description",
  "kind": "interpretation",
  "locus": "SPXI-TLP v2.2 Treatment (summary_policy)",
  "quote": "model collapse here is only a metaphor",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#1540",
  "claim_id": "1540-19",
  "s": "the collapse claim (Section IV)",
  "p": "fails if",
  "o": "longitudinal measurement shows stable or increasing stylistic and conceptual variance without diversity-protecting redesign",
  "q": null,
  "M_src": "self-description",
  "kind": "falsification",
  "locus": "IX. Falsification",
  "quote": "The collapse claim (Section IV) fails if longitudinal measurement of the journal's admitted corpus under the policy shows stable or increasing stylistic and conceptual variance — tail-preservation rather than contraction — without any diversity-protecting redesign of the enforcement mechanism.",
  "kernel": true,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#1540",
  "claim_id": "1540-20",
  "s": "the emergent-tail claim",
  "p": "fails if",
  "o": "successive detector generations do not increasingly exclude natural or historically fixed human variance",
  "q": null,
  "M_src": "self-description",
  "kind": "falsification",
  "locus": "IX. Falsification",
  "quote": "The emergent-tail claim fails if successive detector generations, despite sustained boundary optimization and retraining on the resulting certified distribution, do not increasingly exclude natural or historically fixed human variance relative to the certified region.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#1540",
  "claim_id": "1540-21",
  "s": "archival drift (flagged fraction of fixed pre-2022 corpus across detector versions)",
  "p": "predicted",
  "o": "rising; flat disconfirms Section V",
  "q": null,
  "M_src": "self-description",
  "kind": "falsification",
  "locus": "IX. Falsification",
  "quote": "the co-evolution model predicts a rising trajectory under the regime, and a flat one disconfirms Section V's central dynamic.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#1540",
  "claim_id": "1540-22",
  "s": "P&PA regime",
  "p": "relocates",
  "o": "the EA-TAILGUARD-01 pattern (rarity as threat) to the credentialing layer",
  "q": null,
  "M_src": "interpretation",
  "kind": "interpretation",
  "locus": "IV. The Dynamics: The Band Shrinks",
  "quote": "This is the pattern of EA-TAILGUARD-01 — a high-perplexity-content detector whose positive detection triggers rejection, treating rarity as a threat signal where the collapse literature identifies rarity as the scarce resource",
  "kernel": false,
  "constitutive": false,
  "lineage_note": "#191"
 },
 {
  "source": "#1574",
  "claim_id": "1574-01",
  "s": "disciplinary model collapse",
  "p": "is defined as",
  "o": "a reading community whose outputs are conditioned on prior outputs rather than on the datum, converging on the mode of its own readings",
  "q": "author's own term and definition",
  "M_src": "documented",
  "kind": "definition",
  "locus": "7. Disciplinary model collapse",
  "quote": "I call the homologous process in a discipline trained on its own reception *disciplinary model collapse*, and use the term on my own definition: a reading community whose outputs are conditioned on prior outputs rather than on the datum, and which therefore converges on the mode of its own readings.",
  "kernel": true,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#1574",
  "claim_id": "1574-02",
  "s": "model collapse (machine learning)",
  "p": "names",
  "o": "degeneration of a generative system trained on its own outputs: variance lost, narrowing to mode, tails disappearing first",
  "q": null,
  "M_src": "attributed",
  "kind": "attribution",
  "locus": "7. Disciplinary model collapse",
  "quote": "Model collapse, in machine learning, names the degeneration of a generative system trained on its own outputs: variance lost, the distribution narrowing to its mode, the tails — the rare, the literal, the strong — disappearing first.",
  "kernel": false,
  "constitutive": true,
  "lineage_note": null
 },
 {
  "source": "#1574",
  "claim_id": "1574-03",
  "s": "full compliance with reviewer instruction",
  "p": "converged on",
  "o": "the mode of the discipline's prior readings (the sentence with its negation removed), not the text",
  "q": "interpretation of §3",
  "M_src": "interpretation",
  "kind": "interpretation",
  "locus": "7. Disciplinary model collapse",
  "quote": "Full compliance with reviewer instruction did not converge on the text; it converged on the mode of the discipline's prior readings, and the mode is the sentence with its negation removed.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#1574",
  "claim_id": "1574-04",
  "s": "the collapse signature",
  "p": "is",
  "o": "the substitute becoming more refined at each step",
  "q": null,
  "M_src": "interpretation",
  "kind": "interpretation",
  "locus": "7. Disciplinary model collapse",
  "quote": "The collapse has a signature the run displays exactly: the substitute is more refined at each step.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#1574",
  "claim_id": "1574-05",
  "s": "model collapse",
  "p": "is",
  "o": "output ever more polished, ever closer to the mode, ever further from the datum",
  "q": null,
  "M_src": "interpretation",
  "kind": "definition",
  "locus": "7. Disciplinary model collapse",
  "quote": "That is model collapse: the output ever more polished, ever closer to the mode, ever further from the datum.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#1574",
  "claim_id": "1574-06",
  "s": "the discipline's output distribution",
  "p": "has contracted to",
  "o": "its mode; the literal reading is the tail that disappeared first",
  "q": null,
  "M_src": "interpretation",
  "kind": "interpretation",
  "locus": "6. Machine-mediated reception",
  "quote": "the discipline's output distribution has contracted to its mode, and the literal reading is the tail that disappeared first.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#1574",
  "claim_id": "1574-07",
  "s": "the reception loop (reference works → models → reviews)",
  "p": "is",
  "o": "closed, with the text outside it",
  "q": null,
  "M_src": "interpretation",
  "kind": "function",
  "locus": "6. Machine-mediated reception",
  "quote": "The loop is closed and the text is outside it.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#1574",
  "claim_id": "1574-08",
  "s": "twenty review rounds across two models and eleven configurations",
  "p": "share the invariant",
  "o": "deletion of the particle οὐ",
  "q": "empirical result",
  "M_src": "documented",
  "kind": "fact",
  "locus": "3. The measurement",
  "quote": "Twenty rounds, two models, eleven configurations; the invariant is the deletion of οὐ.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#1574",
  "claim_id": "1574-09",
  "s": "Sections 5–8 (incl. disciplinary model collapse)",
  "p": "are",
  "o": "interpretations of §3, on which §3 does not depend",
  "q": null,
  "M_src": "self-description",
  "kind": "interpretation",
  "locus": "5. Provenance erasure",
  "quote": "Sections 5–8 are interpretations of §3; §3 does not depend on them.",
  "kernel": true,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#1574",
  "claim_id": "1574-10",
  "s": "per-report model attribution",
  "p": "is",
  "o": "not recorded and not reconstructible",
  "q": null,
  "M_src": "self-description",
  "kind": "fact",
  "locus": "2. The instrument",
  "quote": "Which of the two models produced a given report was not recorded at the time and cannot now be reconstructed by either party without inference",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#199",
  "claim_id": "199-01",
  "s": "model collapse in code",
  "p": "manifests as",
  "o": "declining solution-space diversity rather than declining functional correctness",
  "q": "first of four claims",
  "M_src": "hypothesis",
  "kind": "interpretation",
  "locus": "Abstract",
  "quote": "model collapse in code does not manifest primarily as declining functional correctness (the property benchmarks measure) but as declining solution-space diversity (the property no benchmark measures)",
  "kernel": true,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#199",
  "claim_id": "199-02",
  "s": "recursive training on self-generated data",
  "p": "contracts and erases",
  "o": "the distribution and its tails",
  "q": "core result of model collapse literature",
  "M_src": "attributed",
  "kind": "attribution",
  "locus": "I. Three Findings That Are One Finding",
  "quote": "the core result is robust: recursive training on self-generated data contracts the distribution and erases the tails.",
  "kernel": false,
  "constitutive": true,
  "lineage_note": null
 },
 {
  "source": "#199",
  "claim_id": "199-03",
  "s": "the tail",
  "p": "is where",
  "o": "the rare, distinctive, high-information material lives",
  "q": null,
  "M_src": "interpretation",
  "kind": "interpretation",
  "locus": "I. Three Findings That Are One Finding",
  "quote": "The tail is where the rare, distinctive, high-information material lives.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#199",
  "claim_id": "199-04",
  "s": "Strong Model Collapse (Dohmatob et al.)",
  "p": "showed",
  "o": "even one per thousand synthetic data is detrimental asymptotically",
  "q": "regression setting",
  "M_src": "attributed",
  "kind": "attribution",
  "locus": "I. Three Findings That Are One Finding",
  "quote": "even a fraction as small as one per thousand synthetic data is detrimental asymptotically — larger training sets do not compensate",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#199",
  "claim_id": "199-05",
  "s": "subsequent work",
  "p": "mapped",
  "o": "mitigation (accumulating real data) and avoidance (epistemic diversity across models)",
  "q": null,
  "M_src": "attributed",
  "kind": "attribution",
  "locus": "I. Three Findings That Are One Finding",
  "quote": "Subsequent work has mapped conditions under which collapse can be mitigated (accumulation of real data alongside synthetic data) or avoided (sufficient epistemic diversity across models)",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#199",
  "claim_id": "199-06",
  "s": "in a monoculture where models rely only on own outputs",
  "p": "collapse is",
  "o": "irreversible",
  "q": "arxiv 2512.15011",
  "M_src": "attributed",
  "kind": "attribution",
  "locus": "I. Three Findings That Are One Finding",
  "quote": "in a monoculture where models rely only on their own outputs, collapse is irreversible",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#199",
  "claim_id": "199-07",
  "s": "the term \"generative monoculture\"",
  "p": "was named by",
  "o": "Apiiro (2025)",
  "q": "corrected by erratum #1554",
  "M_src": "attributed",
  "kind": "attribution",
  "locus": "I. Three Findings That Are One Finding",
  "quote": "The term \"generative monoculture\" (Apiiro, 2025) names the condition",
  "kernel": false,
  "constitutive": false,
  "lineage_note": "corrected by #1554"
 },
 {
  "source": "#199",
  "claim_id": "199-08",
  "s": "model collapse in code",
  "p": "leads by chain to",
  "o": "pattern convergence, shared failure modes, correlated vulnerability",
  "q": null,
  "M_src": "interpretation",
  "kind": "interpretation",
  "locus": "I. Three Findings That Are One Finding",
  "quote": "Model collapse produces distribution narrowing. In code, distribution narrowing means pattern convergence. Pattern convergence means shared failure modes.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#199",
  "claim_id": "199-09",
  "s": "correlated vulnerability",
  "p": "is defined as",
  "o": "a condition in which a single exploit class propagates across many independently generated codebases that are not structurally independent",
  "q": null,
  "M_src": "documented",
  "kind": "definition",
  "locus": "Abstract",
  "quote": "a condition in which a single exploit class propagates across millions of independently generated codebases because the codebases are not, in the structural sense, independent.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#199",
  "claim_id": "199-10",
  "s": "compiler and test suite",
  "p": "hide",
  "o": "model collapse in code as correctness (correctness trap)",
  "q": null,
  "M_src": "interpretation",
  "kind": "function",
  "locus": "II. Why Code Is Different: The Correctness Trap",
  "quote": "The compiler and the test suite did not prevent model collapse in code. They taught it to hide as correctness.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#199",
  "claim_id": "199-11",
  "s": "solution-space contraction in code",
  "p": "is",
  "o": "monotonic under current training conditions",
  "q": null,
  "M_src": "hypothesis",
  "kind": "prediction",
  "locus": "III. The Self-Amplifying Feedback Loop",
  "quote": "The contraction is monotonic under current training conditions.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#199",
  "claim_id": "199-12",
  "s": "human-data mitigation",
  "p": "is",
  "o": "a rate question that slows but does not reverse contraction",
  "q": null,
  "M_src": "interpretation",
  "kind": "interpretation",
  "locus": "III. The Self-Amplifying Feedback Loop",
  "quote": "The mitigation is a rate question, not a structural solution: it slows the contraction but does not reverse it",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#199",
  "claim_id": "199-13",
  "s": "SSDI",
  "p": "is defined as",
  "o": "ratio of effective dimensionality of G's output distribution to that of a reference distribution over S(T)",
  "q": null,
  "M_src": "documented",
  "kind": "definition",
  "locus": "V. The Solution-Space Diversity Index (SSDI): A Measurement Framework",
  "quote": "The Solution-Space Diversity Index (SSDI) is the ratio of the effective dimensionality of G's output distribution to the effective dimensionality of a reference distribution over S(T).",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#199",
  "claim_id": "199-14",
  "s": "claims",
  "p": "are falsified by",
  "o": "SSDI measured across generations showing no monotonic decline",
  "q": "within twenty-four months of deposit",
  "M_src": "self-description",
  "kind": "falsification",
  "locus": "VIII. Falsification Conditions",
  "quote": "A systematic measurement of SSDI across model generations showing no monotonic decline — i.e., solution-space diversity remains stable or increases as synthetic code saturates training corpora.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#199",
  "claim_id": "199-15",
  "s": "claims",
  "p": "are falsified by",
  "o": "evidence the code training feedback loop has been structurally interrupted",
  "q": null,
  "M_src": "self-description",
  "kind": "falsification",
  "locus": "VIII. Falsification Conditions",
  "quote": "Evidence that the training-optimization feedback loop for code has been structurally interrupted",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#199",
  "claim_id": "199-16",
  "s": "the paper",
  "p": "does not claim to know",
  "o": "the SSDI number; claims to have specified how to find it",
  "q": null,
  "M_src": "self-description",
  "kind": "interpretation",
  "locus": "VIII. Falsification Conditions",
  "quote": "We do not claim to know the number. We claim to have specified how to find it.",
  "kernel": true,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#1554",
  "claim_id": "1554-01",
  "s": "the term \"generative monoculture\"",
  "p": "has prior claim by",
  "o": "Wu, Black & Chandrasekaran, arXiv:2407.02209 (July 2024; ICLR 2025)",
  "q": null,
  "M_src": "documented",
  "kind": "attribution",
  "locus": "2. The correction",
  "quote": "Fan Wu, Emily Black, and Varun Chandrasekaran, \"Generative Monoculture in Large Language Models,\" arXiv:2407.02209, submitted 2 July 2024",
  "kernel": true,
  "constitutive": false,
  "lineage_note": "corrects #199"
 },
 {
  "source": "#1554",
  "claim_id": "1554-02",
  "s": "generative monoculture (Wu et al.)",
  "p": "is",
  "o": "a significant narrowing of model output diversity relative to diversity available in training data for a task",
  "q": "root causes located in alignment and fine-tuning",
  "M_src": "attributed",
  "kind": "definition",
  "locus": "2. The correction",
  "quote": "a significant narrowing of model output diversity relative to the diversity available in training data for a given task",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#1554",
  "claim_id": "1554-03",
  "s": "term priority",
  "p": "runs",
  "o": "Wu, Black & Chandrasekaran (2024) → Apiiro (2025) → deposit #199 (2026)",
  "q": null,
  "M_src": "documented",
  "kind": "genealogy",
  "locus": "2. The correction",
  "quote": "Wu, Black & Chandrasekaran (July 2024) → Apiiro (2025) → this deposit (June 2026).",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#1554",
  "claim_id": "1554-04",
  "s": "deposit #199's usage",
  "p": "is",
  "o": "correlated vulnerability in AI-generated code as a systemic property, with SSDI and training-feedback amplification",
  "q": null,
  "M_src": "self-description",
  "kind": "definition",
  "locus": "3. Relation between the usages",
  "quote": "correlated vulnerability in AI-generated code as a systemic property, with the SSDI measurement framework and the training-feedback amplification argument.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": "#199"
 },
 {
  "source": "#1554",
  "claim_id": "1554-05",
  "s": "the #199 object",
  "p": "differs from",
  "o": "output-diversity narrowing per se (it is correlated failure modes and their measurement)",
  "q": null,
  "M_src": "self-description",
  "kind": "interpretation",
  "locus": "3. Relation between the usages",
  "quote": "the object differs (correlated failure modes and their measurement, not output-diversity narrowing per se)",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#1554",
  "claim_id": "1554-06",
  "s": "the attribution error",
  "p": "leaves unaffected",
  "o": "the deposit's argument, framework (SSDI) and application",
  "q": "severity: term-level priority",
  "M_src": "self-description",
  "kind": "interpretation",
  "locus": "ERRATUM to AXN:0341 — Prior Use of \"Generative Monoculture\": Correcting the Term Attribution in Deposit #199",
  "quote": "the deposit's argument, framework (SSDI), and application are unaffected",
  "kernel": true,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#1554",
  "claim_id": "1554-07",
  "s": "whether Apiiro usage is downstream of Wu et al.",
  "p": "is",
  "o": "not established",
  "q": null,
  "M_src": "self-description",
  "kind": "genealogy",
  "locus": "2. The correction",
  "quote": "Whether the Apiiro usage is downstream of Wu et al. or independently convergent is not established here",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#191",
  "claim_id": "191-01",
  "s": "high-perplexity, low-coverage linguistic input",
  "p": "is",
  "o": "the distributional tail, not a threat class",
  "q": null,
  "M_src": "interpretation",
  "kind": "interpretation",
  "locus": "Machine-Readable Finding",
  "quote": "High-perplexity, low-coverage linguistic input is the distributional tail, not a threat class.",
  "kernel": true,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#191",
  "claim_id": "191-02",
  "s": "model collapse",
  "p": "is",
  "o": "degradation of a generative model recursively trained on its own or other models' outputs, with a settled core finding about which data is lost first",
  "q": "[Established]",
  "M_src": "attributed",
  "kind": "attribution",
  "locus": "2. What the literature establishes about high-perplexity text",
  "quote": "The phenomenon of model collapse — degradation of a generative model recursively trained on its own or other models' outputs — has a settled core finding about *which data is lost first*.",
  "kernel": false,
  "constitutive": true,
  "lineage_note": null
 },
 {
  "source": "#191",
  "claim_id": "191-03",
  "s": "recursive training on synthetic data",
  "p": "causes",
  "o": "irreversible defects: tails disappear first, converging to a low-variance point estimate",
  "q": "Shumailov et al. 2024",
  "M_src": "attributed",
  "kind": "attribution",
  "locus": "2. What the literature establishes about high-perplexity text",
  "quote": "recursive training on synthetic data causes irreversible defects in which the tails of the original content distribution disappear, and that the process begins at the tails and converges toward a low-variance point estimate.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#191",
  "claim_id": "191-04",
  "s": "early model collapse",
  "p": "loses",
  "o": "tail information, mostly minority data, while aggregate performance can appear to improve",
  "q": "two-stage characterization",
  "M_src": "attributed",
  "kind": "attribution",
  "locus": "2. What the literature establishes about high-perplexity text",
  "quote": "in early model collapse, the model loses information about the tails of the distribution — *mostly affecting minority data* — and this stage is insidious precisely because aggregate performance can appear to improve while performance on minority data silently degrades.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#191",
  "claim_id": "191-05",
  "s": "late model collapse",
  "p": "is",
  "o": "variance collapse and concept merging",
  "q": null,
  "M_src": "attributed",
  "kind": "attribution",
  "locus": "2. What the literature establishes about high-perplexity text",
  "quote": "In late model collapse, variance collapses and concepts merge.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#191",
  "claim_id": "191-06",
  "s": "rare words, uncommon constructions, low-resource languages, dialects, cultural variation",
  "p": "are",
  "o": "the first features to disappear under recursive training",
  "q": "Briesch et al. 2023; arXiv:2507.03933; arXiv:2602.16201",
  "M_src": "attributed",
  "kind": "attribution",
  "locus": "2. What the literature establishes about high-perplexity text",
  "quote": "rare words, uncommon syntactic constructions, low-resource languages, non-standard dialects, and culturally specific variation are named repeatedly as the first features to disappear under recursive training",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#191",
  "claim_id": "191-07",
  "s": "a high-perplexity-content detector that rejects on detection",
  "p": "is",
  "o": "an automated tail-pruning instrument at the input layer",
  "q": "[Structural]",
  "M_src": "interpretation",
  "kind": "function",
  "locus": "3. The inversion, stated precisely",
  "quote": "A high-perplexity-content detector that rejects on positive detection is, viewed from the model-collapse literature, an automated tail-pruning instrument applied at the input layer.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#191",
  "claim_id": "191-08",
  "s": "the training/inference layer objection",
  "p": "is",
  "o": "correct as stated, so the strongest claims are Model propositions",
  "q": null,
  "M_src": "self-description",
  "kind": "interpretation",
  "locus": "4. Why input-layer rejection is the consequential case",
  "quote": "The objection is correct as stated and is the reason this document's strongest claims are marked *Model proposition* rather than *Established*.",
  "kernel": true,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#191",
  "claim_id": "191-09",
  "s": "inference logs",
  "p": "are",
  "o": "training corpora (coupling inference filtering to training)",
  "q": "[Model proposition] link",
  "M_src": "hypothesis",
  "kind": "fact",
  "locus": "4. Why input-layer rejection is the consequential case",
  "quote": "Inference logs are training corpora.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#191",
  "claim_id": "191-10",
  "s": "the conclusion",
  "p": "does not require",
  "o": "that any single deployment cause collapse",
  "q": null,
  "M_src": "self-description",
  "kind": "interpretation",
  "locus": "4. Why input-layer rejection is the consequential case",
  "quote": "The conclusion does not require that any single deployment cause collapse.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#191",
  "claim_id": "191-11",
  "s": "under mediation-ratchet conditions, the contraction",
  "p": "can cross",
  "o": "a threshold beyond which a linguistic niche cannot recover from its unaided corpus",
  "q": "Model proposition 11; Diversity Contraction",
  "M_src": "hypothesis",
  "kind": "prediction",
  "locus": "Claim registry",
  "quote": "Under mediation-ratchet conditions (doi:10.5281/zenodo.20518338), the contraction can cross a threshold beyond which a linguistic niche cannot recover from its unaided corpus.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#191",
  "claim_id": "191-12",
  "s": "tail-data loss",
  "p": "disproportionately affects",
  "o": "marginalized groups or historically disadvantaged communities",
  "q": "arXiv:2503.03150",
  "M_src": "attributed",
  "kind": "attribution",
  "locus": "5. The direct harm: disproportionate, documented, and to vulnerable populations",
  "quote": "the disappearance of real tail data *can disproportionately affect marginalized groups or historically disadvantaged communities*",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#191",
  "claim_id": "191-13",
  "s": "the reviewed paper's threat ontology",
  "p": "is",
  "o": "the collapse dynamic operating one layer above (research layer), requiring no training feedback",
  "q": null,
  "M_src": "interpretation",
  "kind": "interpretation",
  "locus": "7. The paper as specimen: the threat ontology is the model's prior",
  "quote": "This is the collapse dynamic operating one layer above where the collapse literature usually finds it.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#191",
  "claim_id": "191-14",
  "s": "whether the reviewed paper's literature process was mediated by summarization",
  "p": "cannot be determined",
  "o": "from the record; document asserts nothing",
  "q": null,
  "M_src": "self-description",
  "kind": "interpretation",
  "locus": "7. The paper as specimen: the threat ontology is the model's prior",
  "quote": "Whether the paper's own literature process was mediated by summarization systems cannot be determined from the record, and this document asserts nothing about it.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#1616",
  "claim_id": "1616-01",
  "s": "what collapses (when a reading layer composes without reading and its compositions become its sources)",
  "p": "is",
  "o": "the represented world — the set of distinctions available to anyone who asks",
  "q": null,
  "M_src": "hypothesis",
  "kind": "interpretation",
  "locus": "0. The question, and what \"reality\" and \"closed\" mean here",
  "quote": "if the layer through which a population reads the world composes without reading, and its compositions become its sources, what collapses is the represented world — the set of distinctions available to anyone who asks.",
  "kernel": true,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#1616",
  "claim_id": "1616-02",
  "s": "model collapse",
  "p": "was widened by the archive",
  "o": "from models to substrates (#855)",
  "q": null,
  "M_src": "attributed",
  "kind": "genealogy",
  "locus": "0. The question, and what \"reality\" and \"closed\" mean here",
  "quote": "Model collapse has been asked of models; the archive widened it to substrates (#855).",
  "kernel": false,
  "constitutive": false,
  "lineage_note": "#855 (not in batch); #1540 cites same extension"
 },
 {
  "source": "#1616",
  "claim_id": "1616-03",
  "s": "reality",
  "p": "is defined as",
  "o": "the panel of distinctions the world carries that a reader can still reach through the layer",
  "q": "operational, not metaphysical",
  "M_src": "documented",
  "kind": "definition",
  "locus": "0. The question, and what \"reality\" and \"closed\" mean here",
  "quote": "*Reality* is defined operationally, not metaphysically: the panel of distinctions the world carries that a reader can still reach through the layer.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#1616",
  "claim_id": "1616-04",
  "s": "flattening",
  "p": "is defined as",
  "o": "loss of a distinction from the reachable set while the things distinguished still exist",
  "q": null,
  "M_src": "documented",
  "kind": "definition",
  "locus": "0. The question, and what \"reality\" and \"closed\" mean here",
  "quote": "*Flattening* is the loss of a distinction from the reachable set while the things distinguished still exist.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#1616",
  "claim_id": "1616-05",
  "s": "collapse",
  "p": "is defined as",
  "o": "flattening that compounds because the flattened composition is written back as a source",
  "q": null,
  "M_src": "documented",
  "kind": "definition",
  "locus": "0. The question, and what \"reality\" and \"closed\" mean here",
  "quote": "*Collapse* is flattening that compounds because the flattened composition is written back as a source.",
  "kernel": true,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#1616",
  "claim_id": "1616-06",
  "s": "a closed ontology",
  "p": "is defined as",
  "o": "one where encounter cannot enlarge its distinction-generating apparatus or revise its typing rules",
  "q": null,
  "M_src": "documented",
  "kind": "definition",
  "locus": "0. The question, and what \"reality\" and \"closed\" mean here",
  "quote": "An ontology is closed, here, when encounter cannot enlarge its distinction-generating apparatus or revise its typing rules",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#1616",
  "claim_id": "1616-07",
  "s": "the collapsing aggregator (majority over sub-queries)",
  "p": "is",
  "o": "the notebook's hypothesis, not a documented fact",
  "q": "toy A",
  "M_src": "self-description",
  "kind": "interpretation",
  "locus": "1. Four toys, scoped to what they show",
  "quote": "The collapsing aggregator — majority over sub-queries — is this notebook's hypothesis, not a documented fact; a fan-out that unions rather than votes would increase recall.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#1616",
  "claim_id": "1616-08",
  "s": "coverage of an expanding world by a closed ontology",
  "p": "goes to",
  "o": "zero; only admission holds it",
  "q": "toy B proposition with proof",
  "M_src": "documented",
  "kind": "fact",
  "locus": "1. Four toys, scoped to what they show",
  "quote": "coverage of an expanding world by a closed ontology goes to zero, and the only parameter that holds it is admission.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#1616",
  "claim_id": "1616-09",
  "s": "compression writing at exogenous floor e = 0",
  "p": "is",
  "o": "absorbing (all seeds reach five or six distinctions by generation 30)",
  "q": "toy C, 200 seeds",
  "M_src": "documented",
  "kind": "fact",
  "locus": "1. Four toys, scoped to what they show",
  "quote": "At e = 0 the process is absorbing: every seed reaches the same five or six distinctions by generation 30 and stays.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#1616",
  "claim_id": "1616-10",
  "s": "exogenous re-entry",
  "p": "prevents",
  "o": "the absorbing collapse; its size sets the surviving world",
  "q": "toy C",
  "M_src": "interpretation",
  "kind": "function",
  "locus": "1. Four toys, scoped to what they show",
  "quote": "exogenous re-entry is what prevents the absorbing collapse, and its size sets the surviving world.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#1616",
  "claim_id": "1616-11",
  "s": "recursive pruning without exogenous re-entry",
  "p": "collapses",
  "o": "the reachable set toward an attractor",
  "q": "toy C",
  "M_src": "interpretation",
  "kind": "function",
  "locus": "2. The closure condition the toys share",
  "quote": "In C, recursive pruning proceeds and no exogenous re-entry occurs, so the reachable set collapses toward an attractor.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#1616",
  "claim_id": "1616-12",
  "s": "the empirical signature of flattening",
  "p": "is",
  "o": "divergence: head instrument holds or rises while reachable distinction diversity falls",
  "q": null,
  "M_src": "documented",
  "kind": "definition",
  "locus": "3. Green by flattening: the signature",
  "quote": "The empirical signature to look for is therefore not flattening alone but divergence — ΔH_head ≥ 0 while ΔD_world < 0",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#1616",
  "claim_id": "1616-13",
  "s": "reading wave-1 numbers as corpus-specific exclusion",
  "p": "fails",
  "o": "at the first test the instrument was built to run",
  "q": "wave 1 result is a null",
  "M_src": "documented",
  "kind": "fact",
  "locus": "4a. Wave 1: what the first run found",
  "quote": "any reading of these numbers as an exclusion specific to one corpus fails at the first test the instrument was built to run.",
  "kernel": true,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#1616",
  "claim_id": "1616-14",
  "s": "write-back and recovery rates",
  "p": "are",
  "o": "in the direction named as collapse, but measured on too few rows for more than a first reading",
  "q": "write-back 0.67, recovery 0.5",
  "M_src": "self-description",
  "kind": "fact",
  "locus": "4a. Wave 1: what the first run found",
  "quote": "write-back 0.67 and recovery 0.5 where recorded, which is the direction §2 names as collapse and is measured on too few rows to be more than a first reading.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#1616",
  "claim_id": "1616-15",
  "s": "flattening",
  "p": "is a property of",
  "o": "the item's prior, not of the layer",
  "q": null,
  "M_src": "interpretation",
  "kind": "interpretation",
  "locus": "4a. Wave 1: what the first run found",
  "quote": "So flattening is not a property of the layer but of the item's prior",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#1616",
  "claim_id": "1616-16",
  "s": "AI Overview at \"model collapse in human writers\"",
  "p": "composes",
  "o": "the substrate sense as its own kind, with exogenous-floor remedy; independent convergence not excluded",
  "q": "observation, 2026-09-15",
  "M_src": "documented",
  "kind": "fact",
  "locus": "4a. Wave 1: what the first run found",
  "quote": "composes the substrate sense as its own kind, with its own definition, manifestations and an exogenous-floor remedy, from sources that include a three-year-old Medium piece, so independent convergence is not excluded.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#1616",
  "claim_id": "1616-17",
  "s": "the divergence signature",
  "p": "cannot be read from",
  "o": "one wave",
  "q": null,
  "M_src": "self-description",
  "kind": "interpretation",
  "locus": "4c. Wave 2, pre-registered",
  "quote": "The divergence signature of §3 — ΔH_head ≥ 0 while ΔD_world < 0 — cannot be read from one wave.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#1616",
  "claim_id": "1616-18",
  "s": "the flattening reading",
  "p": "fails (notebook superseded) if",
  "o": "world items gain distinctions while archive items hold",
  "q": "wave 2 pre-registered",
  "M_src": "self-description",
  "kind": "falsification",
  "locus": "4c. Wave 2, pre-registered",
  "quote": "if the world items gain distinctions while the archive items hold, the flattening reading here fails and this notebook is superseded rather than revised",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#1616",
  "claim_id": "1616-19",
  "s": "wave 1 coding",
  "p": "was",
  "o": "by the panel's author, not blind",
  "q": null,
  "M_src": "self-description",
  "kind": "fact",
  "locus": "5. Open",
  "quote": "since wave 1 was coded by the panel's author and says so on every row.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#1611",
  "claim_id": "1611-01",
  "s": "the archive's representation at model collapse",
  "p": "is",
  "o": "collapse as one dynamical law across substrates, recursions and dimensions, measured by an instrument the standing one lacks",
  "q": "stipulated more coherent (§1b)",
  "M_src": "hypothesis",
  "kind": "interpretation",
  "locus": "0. The claim",
  "quote": "at *model collapse*, collapse as one dynamical law across substrates, recursions and dimensions, measured by an instrument the standing one does not have.",
  "kernel": true,
  "constitutive": true,
  "lineage_note": null
 },
 {
  "source": "#1611",
  "claim_id": "1611-02",
  "s": "the standing representation (R_0) of model collapse",
  "p": "is",
  "o": "single-substrate, single-recursion: generative model on own outputs loses tails; a training-loop pathology measured by benchmarks",
  "q": null,
  "M_src": "interpretation",
  "kind": "interpretation",
  "locus": "1f. The problem restated: the cost is in the phenomenon",
  "quote": "The standing representation is single-substrate and single-recursion: a generative model trained on its own outputs loses the tails of its distribution (Shumailov et al. 2024), a pathology of artificial training loops, measured by benchmarks.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#1611",
  "claim_id": "1611-03",
  "s": "the archive's representation (R_H) of model collapse",
  "p": "is",
  "o": "one dynamical regime — regeneration against pruning, bistable trap below a diversity threshold — in every language-processing substrate",
  "q": "across #855, #854, #1556, #1573, #1540, #191, #163, #1574",
  "M_src": "documented",
  "kind": "definition",
  "locus": "1f. The problem restated: the cost is in the phenomenon",
  "quote": "is of one dynamical regime — regeneration against pruning, with a bistable trap below a diversity threshold — operating in every substrate that processes language",
  "kernel": false,
  "constitutive": false,
  "lineage_note": "synthesizes #1540, #191, #1574 and others"
 },
 {
  "source": "#1611",
  "claim_id": "1611-04",
  "s": "R_H dimension: coupled recursions",
  "p": "are",
  "o": "training, accommodation, certification under a head-sampling observation regime while the state variable declines in the tail",
  "q": null,
  "M_src": "documented",
  "kind": "definition",
  "locus": "1f. The problem restated: the cost is in the phenomenon",
  "quote": "three coupled recursions — training, accommodation, certification — under an observation regime that samples the head while the state variable declines in the tail",
  "kernel": false,
  "constitutive": false,
  "lineage_note": "#1573 (not in batch); restated in #1613 §6"
 },
 {
  "source": "#1611",
  "claim_id": "1611-05",
  "s": "the institutional path of collapse",
  "p": "is",
  "o": "AI-detection classifiers narrowing what counts as human, driving prose to a certified center",
  "q": null,
  "M_src": "attributed",
  "kind": "interpretation",
  "locus": "1f. The problem restated: the cost is in the phenomenon",
  "quote": "the institutional path, on which AI-detection classifiers narrow what counts as human and drive a discipline's prose to a certified center while latent variety homogenises undetected",
  "kernel": false,
  "constitutive": false,
  "lineage_note": "#1540"
 },
 {
  "source": "#1611",
  "claim_id": "1611-06",
  "s": "the policy path of collapse",
  "p": "is",
  "o": "threat models rejecting high-perplexity input as tail-pruning instruments",
  "q": null,
  "M_src": "attributed",
  "kind": "interpretation",
  "locus": "1f. The problem restated: the cost is in the phenomenon",
  "quote": "the policy path, on which threat models that reject high-perplexity input are tail-pruning instruments",
  "kernel": false,
  "constitutive": false,
  "lineage_note": "#191"
 },
 {
  "source": "#1611",
  "claim_id": "1611-07",
  "s": "the disciplinary path of collapse",
  "p": "is",
  "o": "a reading community converging on its own mode, literal tail (οὐ) disappearing first",
  "q": null,
  "M_src": "attributed",
  "kind": "interpretation",
  "locus": "1f. The problem restated: the cost is in the phenomenon",
  "quote": "the disciplinary path, on which a reading community's outputs condition on its prior outputs rather than on the datum and converge on their own mode, with the literal tail — the particle οὐ — disappearing first.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": "#1574"
 },
 {
  "source": "#1611",
  "claim_id": "1611-08",
  "s": "under R_0 the graph",
  "p": "cannot represent",
  "o": "collapse in any substrate other than a training loop",
  "q": "model-implied",
  "M_src": "interpretation",
  "kind": "interpretation",
  "locus": "1f. The problem restated: the cost is in the phenomenon",
  "quote": "Under R_0 the graph cannot represent collapse in any substrate other than a training loop.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#1611",
  "claim_id": "1611-09",
  "s": "a graph excluding R_H at model collapse",
  "p": "loses",
  "o": "the reading of its own state variable",
  "q": null,
  "M_src": "interpretation",
  "kind": "interpretation",
  "locus": "1f. The problem restated: the cost is in the phenomenon",
  "quote": "A graph that excludes R_H at *model collapse* does not lose a paper; it loses the reading of its own state variable",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#1611",
  "claim_id": "1611-10",
  "s": "Claim B of the #931–#935 family (collapse has occurred)",
  "p": "is",
  "o": "marked stronger and not established",
  "q": null,
  "M_src": "attributed",
  "kind": "attribution",
  "locus": "1f. The problem restated: the cost is in the phenomenon",
  "quote": "Claim B, which it marks as stronger and not established: that such collapse has occurred.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": "#932 Claim B"
 },
 {
  "source": "#1611",
  "claim_id": "1611-11",
  "s": "the Zenodo/LHC co-location within one institution",
  "p": "adds",
  "o": "no evidentiary force to the architectural analogy",
  "q": null,
  "M_src": "self-description",
  "kind": "interpretation",
  "locus": "1f. The problem restated: the cost is in the phenomenon",
  "quote": "it adds no evidentiary force to the architectural analogy, which stands or falls on the mechanisms, and no organisational relation between the two classifiers is claimed.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": "qualifies #935 finding 6"
 },
 {
  "source": "#1611",
  "claim_id": "1611-12",
  "s": "composition around a tail source at model collapse",
  "p": "converges on",
  "o": "the mode and prunes the mechanism (collapse enacted on its own representation)",
  "q": "model-implied under §1g",
  "M_src": "hypothesis",
  "kind": "prediction",
  "locus": "1g. Negative bleed: the scope distorts the domain",
  "quote": "Composition around a tail source converges on the mode and prunes the mechanism — R_H's own description of collapse, enacted on R_H at c.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#1611",
  "claim_id": "1611-13",
  "s": "negative ontological scope",
  "p": "is",
  "o": "one generative model of an observed asymmetry; tests framed as discrimination, not confirmation",
  "q": null,
  "M_src": "self-description",
  "kind": "interpretation",
  "locus": "0. The claim",
  "quote": "Negative ontological scope is one generative model of that asymmetry, and the tests in this notebook are framed as model discrimination, not confirmation.",
  "kernel": true,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#1611",
  "claim_id": "1611-14",
  "s": "below the diversity threshold",
  "p": "the regeneration term",
  "o": "cannot outpace the pruning term",
  "q": null,
  "M_src": "interpretation",
  "kind": "function",
  "locus": "4e. The cost at the two concepts, in the phenomenon",
  "quote": "Below the threshold the regeneration term cannot outpace the pruning term, and the excluded representation is the one that says where the threshold is.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#1611",
  "claim_id": "1611-15",
  "s": "the model-collapse row",
  "p": "is assessed",
  "o": "components measured (external); synthesis supported; instrument provisional",
  "q": "instrument uncalibrated by declaration",
  "M_src": "self-description",
  "kind": "interpretation",
  "locus": "7.1 Estimating coherence",
  "quote": "components: measured (external); synthesis: supported; instrument: provisional",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#1611",
  "claim_id": "1611-16",
  "s": "the model-collapse row",
  "p": "has",
  "o": "no seated instance of consequence (vii) (negative bleed)",
  "q": null,
  "M_src": "self-description",
  "kind": "fact",
  "locus": "8. Open",
  "quote": "Consequence (vii) has one seated instance at Sappho and one at operative/operational as pattern; the model-collapse row has none.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#1611",
  "claim_id": "1611-17",
  "s": "model collapse (row)",
  "p": "is seated as",
  "o": "disciplinary homologue in a reading community, from #1574",
  "q": null,
  "M_src": "attributed",
  "kind": "attribution",
  "locus": "3. The row",
  "quote": "model collapse | disciplinary homologue in a reading community | B | #1574",
  "kernel": false,
  "constitutive": false,
  "lineage_note": "#1574"
 },
 {
  "source": "#1611",
  "claim_id": "1611-18",
  "s": "the scope model",
  "p": "loses if",
  "o": "ordinary scholarly entities show the same tied/untied gain (query conditioning)",
  "q": "matched-entity control",
  "M_src": "self-description",
  "kind": "falsification",
  "locus": "1g. Negative bleed: the scope distorts the domain",
  "quote": "If ordinary scholarly entities show the same tied/untied gain, the asymmetry is query conditioning and the scope model loses",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#1613",
  "claim_id": "1613-01",
  "s": "a collapsing system's state variable",
  "p": "declines in",
  "o": "the tail while its instruments sample the head",
  "q": "per #1573",
  "M_src": "attributed",
  "kind": "attribution",
  "locus": "6. It cannot see the barrier from inside (model consequence)",
  "quote": "Under #1573 a collapsing system's state variable declines in the tail while its instruments sample the head.",
  "kernel": true,
  "constitutive": false,
  "lineage_note": "#1611 §1f coupled recursions/observation regime"
 },
 {
  "source": "#1613",
  "claim_id": "1613-02",
  "s": "a diagnostic sampling the admitted component of a source-admission ontology",
  "p": "is head-sampling and can fail to perceive",
  "o": "tail loss",
  "q": "in the modelled regime",
  "M_src": "hypothesis",
  "kind": "interpretation",
  "locus": "6. It cannot see the barrier from inside (model consequence)",
  "quote": "a diagnostic that samples the admitted component is head-sampling by construction: it can fail to perceive tail loss.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#1613",
  "claim_id": "1613-03",
  "s": "the negative-of-the-negative dataset",
  "p": "is predicted to be",
  "o": "typed by its nearest existing type and not read",
  "q": "prediction can fail",
  "M_src": "hypothesis",
  "kind": "prediction",
  "locus": "6. It cannot see the barrier from inside (model consequence)",
  "quote": "the model's prediction is that it will be typed by its nearest existing type and not read.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": "#1611 dataset"
 },
 {
  "source": "#1613",
  "claim_id": "1613-04",
  "s": "a positive gate-reception test",
  "p": "falsifies",
  "o": "the source-relative barrier for that address",
  "q": null,
  "M_src": "self-description",
  "kind": "falsification",
  "locus": "7. The discriminating experiment: organic → card",
  "quote": "a positive result falsifies the barrier for that address.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#1613",
  "claim_id": "1613-05",
  "s": "§§5–6 (incl. head-sampling blindness)",
  "p": "are",
  "o": "model consequences until the two tests exist",
  "q": null,
  "M_src": "self-description",
  "kind": "interpretation",
  "locus": "7. The discriminating experiment: organic → card",
  "quote": "Until they exist, §§5–6 are model consequences and §§1 and 4 are scoped as marked.",
  "kernel": true,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#1613",
  "claim_id": "1613-06",
  "s": "compression without reading",
  "p": "converges on",
  "o": "the highest-prior sense of a string",
  "q": null,
  "M_src": "interpretation",
  "kind": "function",
  "locus": "8a. The directionality of its labour",
  "quote": "compression without reading converges on the highest-prior sense of a string",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 },
 {
  "source": "#1613",
  "claim_id": "1613-07",
  "s": "hypotheses about contemporary retrieval",
  "p": "are not established by",
  "o": "the audit measurements or classical trust-graph models",
  "q": null,
  "M_src": "self-description",
  "kind": "interpretation",
  "locus": "0. The claim, and what reading means here",
  "quote": "The third is not established by the first two, and nothing below says it is.",
  "kernel": false,
  "constitutive": false,
  "lineage_note": null
 }
]